NAVS
← Back to blog

How to Try On Clothes with ChatGPT: A Step-by-Step Guide to AI Virtual Try-On

N
Navs
Published on October 7, 20266 min read
Tags:ChatGPTtutorialAI shoppingvirtual try-onOpenAI
How to Try On Clothes with ChatGPT: A Step-by-Step Guide to AI Virtual Try-On

Shopping for clothes online is mostly guesswork. You study a size chart, squint at a model's photo, and hope the jacket looks the same on you. On October 1, 2026, OpenAI gave ChatGPT a more useful trick: a virtual try-on button that renders clothing and accessories directly on your own photo, right inside the chat. This guide walks you through exactly how to use it, what to prepare, and where its limits are — so you can decide whether it's worth a spot in your shopping routine.

ChatGPT screenshot

What Launched

OpenAI announced the global launch of two shopping features in ChatGPT on October 1, 2026. The first is a virtual try-on: a "Try on" button that appears on product listings for clothes and accessories in ChatGPT's shopping results. Select it, upload a selfie or full-body photo, and ChatGPT generates an image of the item on you. Your reference photo is saved so you can reuse it for future try-ons without uploading it again.

The second feature is Favorites: a way to save products you discover to a Library inside the app, with folders to organize them. Your saved items sit alongside your try-on images, so everything from a shopping session lives in one place. OpenAI says both features are powered by the ChatGPT Images 2.5 model, which the company claims produces more natural lighting and richer textures, follows editing instructions more reliably, and reduces image-generation latency. Those performance claims come from OpenAI; the company has not published independent benchmarks for garment fidelity.

This is OpenAI's second serious swing at commerce inside ChatGPT. An earlier instant-checkout feature did not perform well, and the emphasis has since shifted toward discovery and visualization — helping you decide what to buy rather than processing the purchase. Google launched its own virtual try-on last year, and Pinterest has long occupied the fashion-discovery space, so ChatGPT is entering crowded territory.

What You Need Before You Start

The barrier to entry is low, but a little setup makes the results noticeably better:

  • A ChatGPT account on web or mobile. OpenAI says the features are available on both platforms, with a global rollout — no waitlist.
  • A clear reference photo. Take or choose a well-lit selfie or full-body photo with a plain background. Avoid cluttered backgrounds, hats, and heavy shadows; the model drapes the garment over whatever it sees, so a clean, front-facing photo gives it the least to guess about.
  • Realistic expectations. The output is a visualization, not a fitting. OpenAI's own documentation states that try-on images may not represent the product or your appearance exactly and do not guarantee fit or size. You still need the merchant's measurements, product details, and return policy before you buy.

How to Try On an Item from Shopping Results

This is the main path OpenAI designed for the feature.

  1. Ask ChatGPT to shop for something. Describe what you want in plain language, for example: "Show me navy rain jackets under $150." ChatGPT will return product cards with imagery, details, and merchant links.
  2. Find a clothing or accessory listing with a "Try on" button. The button appears on eligible product listings in the results.
  3. Select "Try on." ChatGPT will prompt you to take or upload a selfie (or a full-body photo).
  4. View your generated try-on image. ChatGPT Images generates a preview of the item on your photo. Your reference photo is saved for future try-ons.
  5. Iterate if needed. If the preview looks off, try a different reference photo with better lighting, or ask ChatGPT for variations. Remember that the button only appears on supported listings — if it's missing, the item may not be eligible for try-on.

How to Try On Any Item You Find Elsewhere

You are not limited to products ChatGPT surfaces. According to OpenAI, you can also upload a picture of clothing or accessories directly in a conversation — for example, a screenshot of an outfit from Instagram or a retailer's website — and ask ChatGPT to show how the item could look on you. This is genuinely useful: it turns any product photo on the internet into a try-on candidate, without waiting for merchants to opt into a fitting feature.

Save What You Like with Favorites

Once you start generating previews, they pile up. Favorites gives you a place to keep them:

  • Save a product from shopping results to add it to your Favorites.
  • Organize finds into folders inside your ChatGPT Library — useful for comparing, say, three jackets side by side before deciding.
  • Manage your reference photos under Settings > Personalization > Reference photos, where you can change or delete the saved photo. Since the photo persists across sessions, decide deliberately whether you want ChatGPT to keep it; deleting it there is the documented control.

Two Bonus Shopping Workflows

Beyond try-on, OpenAI points to two open-ended ways ChatGPT can help you shop:

  • Style building. Describe a look — "a black-tie dress with a halter neckline" — and ask ChatGPT to shop for the pieces needed to complete it.
  • Celebrity-outfit matching. Upload a photo of a celebrity's outfit and ask ChatGPT to find the items they are wearing that are available for purchase.

Neither is a fitting tool; both are discovery tools. Used together with try-on, they form a short loop: get inspired, find the pieces, preview them on yourself, save the winners.

Limitations You Should Know

The feature is honest about what it cannot do, and you should take those limits seriously:

  • No fit guarantee. The image is a visualization. It does not tell you how the fabric drapes, how a size will feel on your body, or whether the item matches its listing exactly. Always verify against the merchant's size chart and reviews.
  • Representation caveats. Generative previews can smooth over real differences in body shape, skin tone, or proportions. Treat the output as a rough draft of the idea, not a portrait of the result.
  • Reference photos are stored. Convenience comes with a small privacy tradeoff: your photo stays saved for future try-ons until you delete it. Manage it in Settings if that matters to you.
  • Availability is still rolling out. As with most OpenAI launches, some users see features before others, and eligible listings vary by product and merchant.

Who Benefits Most

Not every shopper needs an AI fitting room, but three groups get real value from it. Serial online buyers who order three sizes and return two can pre-filter the obvious mismatches — a preview won't tell you the waist is right, but it will tell you the silhouette is wrong. People with uncommon body shapes that catalog models never represent get a more relevant visualization than any product photo, because the reference image is their own. And gift shoppers can check whether a surprise pick looks plausible on a family member's photo before ordering, saving an awkward return conversation later.

The common thread: use try-on as a filter, not a verdict. It removes the clearly wrong options fast, leaving a shortlist you verify the old-fashioned way — measurements, reviews, and return policies.

Is It Worth Using?

For a specific job — narrowing down which of several items is worth ordering, or checking whether a bold pattern suits you before committing — virtual try-on earns its place. It replaces the mental gymnastics of imagining a garment on your body with an actual image, and the favorites library keeps the session's output organized. It does not replace trying things on in a store, reading reviews, or checking return policies. Used that way, as a discovery and shortlisting tool, it's a meaningful upgrade to shopping inside a chatbot.

If you regularly buy clothes online, try it on your next purchase: generate previews for two or three contenders, save the best ones to Favorites, and then — before you click buy — do the unglamorous check of measurements and the return window. The AI can show you the jacket on your shoulders; only the size chart can tell you it will fit.

References

Share this article
Previous PostReflection Beam vs DeepSeek V4: Which Open-Weight Model Should You Bet On?